Technique for Automatic Classification of Meteor Trails and Other Propagation Mechanisms for the Air Force High Latitude Meteor Burst Test Bed

摘要

This report describes a rule-driven artificial intelligence technique for automating the identification of propagation mechanisms and classification of meteor trail types on the USAF high latitude meteor burst test bed. The procedure reduces the time required to process data from one month of test bed operation from 2 staff months to about 12 hours. It is designed to emulate the procedures used by a human operator with the additional advantage of being able to use previous and future data records in the decision process. The performance of the automatic classification procedure is comparable in terms of correct decisions to human classifiers.